Skip to content

Conversation

@jfkthame
Copy link
Owner

@jfkthame jfkthame commented Oct 3, 2019

No description provided.

jfkthame and others added 30 commits January 16, 2017 18:04
As noted by @mujjingun in issue 63 -- at least, the report looks correct to me: fixes #63.
Fix incorrect test reported in glyph_table.cc
sfntly/c++ defines several confusing typedefs for standard types.  I
used sed to change them back to the standard types.  This will make it
easer for someone new to this codebase (i.e. everbody) to hack on it.

Also, the CMakeLists.txt file assumes you use either GCC or MSVC.  I
use Clang, so I modified it to give Clang the same options as GCC.

Also, the unit tests assume that googletest is located in
`cpp/ext/gtest`.  I added this to a .gitignore file so that I can
symlink to `..../googletest/googletest`.
UnicodeString and U_* are used without including any ICU headers. This
used to work with ICU 58.x, but not any more with ICU 59-to-be.

This change is necessary for upgrading Chrome's ICU to 59.x.
( https://crbug.com/699469 )

Will fix #72
Fix a couple more crashes found by sfntly_fuzzer
The fix for NULL pointer derefs in sfntly::Font::Builder is too
aggressive.

Unfixes https://crbug.com/663737
Fixes https://crbug.com/705357
This time, only check a readable hmtx table.

Fixes https://crbug.com/663737
Commit 1fba3b3 caused large ULONG values to be reinterpreted as signed
32-bit integers.
Fix a variety of sfntly bugs
- Simplify iterator usage.
- Use std::numeric_limits in a couple places.
- Remove unused FontFactory::SetSerializationTableOrdering().
- Mark more variables/methods as const.

No behavior changes.
Cleanup random bits of code.
This prevents a potential infinite loop.

Fixes https://crbug.com/712021
Some of them do not properly check bounds when accessing the
contour_index_ array.
Those that access contour_index_ should be within bounds.
Check read value in CompositeGlyph::Initialize()
The conversion from HTML to MarkDown has been done mainly by Pandoc,
with some manual cleanup afterwards.

Fixes #83.
For negative values the >> operator led to an endless loop.

Using the math provided by the Java standard library is safer.
inbreaks and others added 17 commits October 11, 2017 18:33
Validate EBLC table's numSizes field.
When sfntly depends on ICU4C a shared library, the shared
library may not provide stable C++ ABI. Use ICU C API instead.
UChar is not always char16_t in some platforms. In some cases,
it's wchar_t or uint16_t.

Use std::basic_string<UChar> to avoid std::u16string.
In findMatch() stop searching back for a match if we've exceeded the maximum distance that can be encoded.
[eot conversion] Fix eot conversion failure.
Fix an uninitialized value
This will revert to earlier behavior of HasName, by including a space when appending "Regular" to a base font name to synthesize a possible full name. (The space appears to have been inadvertently dropped during the migration to UCharString.)
Restore space before "Regular" font name suffix
@jfkthame jfkthame merged commit a140ff2 into jfkthame:master Oct 3, 2019
@felipesanches felipesanches deleted the master branch March 2, 2021 04:22
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.